the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 Oct 2024
Low-level jets in the North and Baltic Seas: Mesoscale Model Sensitivity and Climatology
Abstract. Low-level jets (LLJs) are wind speed maxima in the lower part of the atmospheric boundary layer. Accurately accounting for these mesoscale phenomena in wind resource assessment is increasingly important as the height of wind turbines continues to grow. During LLJ events, wind speeds increase, leading to a general increase in power output. We utilize wind measurements from LiDARs and a mast for five sites in the North and Baltic Seas to assess the quality of the WRF model simulations for LLJ characterization. We also investigate the benefits of using the WRF model simulations compared to the widely used ERA5 reanalysis. In the WRF model simulations, we vary the grid spacing, vertical resolution, and the planetary boundary layer and land surface schemes, parameters we deemed most likely to have a substantial impact. The model’s performance is evaluated based on its ability to replicate observed distributions of LLJs and relevant associated characteristics, such as the shear and veer across the rotor-plane of typical large offshore wind turbines (30 m to 300 m). Finally, we generate a five-year LLJ climatology based on using the best-evaluated model configuration. The modeling results show a strong dependency of the LLJ representation and the associated wind profiles on the WRF model configuration and that relying on ERA5 for LLJ characterization is insufficient. For example, the LLJ rate-of-occurrence varied by up to a factor of three or more between some WRF model runs. The simulation using the optimized model configuration more accurately reflects the frequency, intensity, and vertical extension of LLJs, as confirmed by LiDAR data. In the North and Baltic Seas, LLJs occur along the western sea basins around 10–15 % of the time, with average jet heights between 140–220 meters, which are well within the height of operation of modern wind turbines. The most LLJ-prone region is east of southern Sweden, especially during spring and summer. These mesoscale phenomena contribute to up to 30 % of the wind capacity in some areas in this season. Analysis of the five-year modeled LLJ climatology in the North and Southern Baltic Seas gives insights into the physical mechanisms that create them.
- Preprint
(25216 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 25 Dec 2024)
-
RC1: 'Comment on egusphere-2024-3123', Anonymous Referee #1, 22 Nov 2024
reply
Review of the manuscript EGUSPHERE-2024-3123
Low-level jets in the North and Baltic Seas: Mesoscale Model Sensitivity and Climatology
By Olsen an co-workers.
Summary: This paper studies the climatology of low-level jets (frequency, intensity, wind maximum height, duration) in observations over the North Sea and the Baltic Sea, and how these are represented in the WRF model, NEWA (wind atlas) and ERA5. For WRF an ensemble of configurations is set up in terms of PBL physics, number of model layers and domain size. It is found that the ERA5 model underestimates LLJ speed and dilutes the LLJ over a too deep layer. There appears to be considerable difference in timing, location and shapes of the LLJ climatology between WRF configurations. This is important information for wind energy resource assessment, and load assessment and for the development of future wind atlases.
Recommendation: revise
Major remarks:
- The abstract is relatively long and reports quite some of the activities that were employed in the study. So now it is more like a summary. However, it would be good to rewrite the abstract such that it reports more about what has been learnt, i.e. focus on the scientific knowledge/insights that were gained from the study.
- I find the paper has quite many figures. On the one hand this means that the paper is very complete, but at the same time it is a less attractive invitation to read the paper. Also many of the figures are not discussed in very much detail (some in 4-5 lines). So it would be great if the paper can be revised such that the discussion is deepened (link model behaviour to atmospheric physics and dynamics, or to modelling strategies like nudging, initialization, time stepping etc etc), and end up with less figures.
- The members of the WRF ensembles do not have very intuitive names/labels, which means for any discussion of a specific member, the reader has to go back to the table where they have been defined. This is unhandy and time consuming. Could you perhaps do an effort to revise the names in a more intuitive way?
- In terms of positioning the paper, one could link it more to earlier wind atlases results. In fact you make kind of a wind atlas with 5 y of WRF simulations. For a wind atlas it is a short time span, but the main sensitivities remain the same if you would have run 30 years. So I think the paper can benefit from being positioned in a storyline of the making of wind atlases, and/or positioned in relation to recent wind atlas products like NEWA, DOWA, … and whether similar biases are seen. For example Kalverla et al (2020, https://rmets.onlinelibrary.wiley.com/doi/full/10.1002/qj.3748) discusses similar diurnal and seasonal cycles of LLJ climatology in three wind atlases.
Minor remarks
Ln 6: we vary -> we perform sensitivity experiments….
Ln 8: replicate -> reproduce
Ln 19: gives insights: what are these insights? Please add to the abstract.
Ln 32: During LLJ events, wind speeds increase, leading to -> LLJ events lead to
Ln 37: I think that in addition to Kalverla et al. (2019), this is also a good reference based on observations over the North Sea: https://www.sciencedirect.com/science/article/pii/S0167610516307061
Ln 67-69: These read as a figure caption. Better not to put in the main text.
Ln 75: Please include some words whether the period is different or not from the typical climatology.
Ln 101: please add the model version of WRF
Ln 115: if the goal of this final set of simulations is to realize a multi-year dataset, then P_3DTKE is not the most intuitive name to facilitate the reader.
Table 3: it is unclear how turbulent diffusion was dealt with. In the text and in Table the paper mentions a variety of PBL schemes, while table 3 suggests Smagorinsky closure was used everywhere. Please clarify.
Ln 143-144: change of the thresholds. Please justify more why this is allowed, and how much it affected your results. I.e. at least mention that the number of samples increased from XXX to YYY cases.
Ln 152: … consistent. Please add a sentence or two how your work is consistent in this context. How do you ensure you do the right thing for answering your research question.
Ln 160: log-linear interpolation. This will not work in the case of a present LLJ as you show in Fig 2. In which % of the cases do you interpolate wind speed profiles with LLJ properties using this method?
Ln 172: This EMD method sounds interesting, but is not used very often (but that does not mean I am not confident in it). Most studies would likely use a Mann-Whitney-U test to test whether the distributions are the same or not. So why is this alternative EMD method preferred over more often used methods? Likely you have a good reason for it, and could be sold as an innovative part of the study.
Ln 182: shear estimation: please add a description how you include negative shear in your estimation? i.e. if you have 1 model layer with du/dz=1 and the layer above du/dz=-1 (like for a LLJ top), is the total shear then 0, or 2?
Ln 190: Is it a problem for the Z score that the wind speed distribution is not normally distributed?
Ln 190: ensemble members. It is the first time here that ensemble members are mentioned, but the reader does not know here how these ensemble is developed.
Figure 4: Can more be said about Figure 4. It now only gets 5 lines, which means maybe the figure is not needed. For example whether the data gaps occur particularly in LLJ favorable seasons or not, and whether that may result in a biased understanding of the LLJs.
Figure 4: the captions should explain what are NS1, NS2, BS1, BS2 and Osterild N.
Figure 5: I don’t know whether it will appear later, but here as a reader I expect some words whether the climatology found is in agreement or disagreement with earlier studies over the North Sea and Baltic Sea. And if so, can you explain these differences.
Figure 5: please add in the caption what is the time step of the data. Ie. whether we look at a frequency of the time in 10 min boxes or 1 h boxes.
Figure 5: Please add whether the time is in UTC or local time in panel b.
Figure 5: Panel b shows a clearly different timing of the peak of the LLJ between the two seas. This needs to be mentioned and explained in the text.
Ln 241: 2ms−1 and 20% falloffs. Why do you deviate here from your more relaxed criterion explained in the Method (1.5 m/s)?
Figure 7: The error bar represents the spread among the five sites. Please be more specific whether the spread is 1 or 2 standard deviations, or is it really the min and the max. In the latter case it is better perhaps to turn into a whisker plot that also contains the quantiles.
Figure 7d and 7e: the range is relatively small, so better start the y axis not at 0.
Figure 8: there are more bars in the bar graphs than are in the legend. Please complete.
Ln 279-280: If Fig 8 only gets one sentence of attention, is it then really needed?
Ln 297: add a ) behind the “on average”
Figure 10: caption “Maps of modelled LLJ occurrence ….”
Figure 12: in the WRF model climatology…. Please add in the caption which WRF ensemble member is used.
Table A1: header “Obuhkov length (L)” -> should be “Obukhov length (L)”
Citation: https://doi.org/10.5194/egusphere-2024-3123-RC1 -
RC2: 'Comment on egusphere-2024-3123', Anonymous Referee #2, 24 Nov 2024
reply
The manuscript focuses on low-level jets over the North Sea and the Baltic Sea. The authors discuss how these are simulated using the WRF model with various setups. They compare their results against observations as well as the NEWA and ERA5 datasets. The analysis is important not only from a scientific perspective but also for wind energy resource assessment and applications, although the latter appears somewhat inconsistent with the overall analysis presented in the paper. The manuscript is well-written and clear. However, it should be considered for publication only after a revision is undertaken.
Major Remarks:
- The analysis appears to have been conducted in sections that seem somewhat disconnected from each other. It is unclear why the authors selected different setups for the E and S simulations compared to the P simulations. Why not use the P domain with a 3 km resolution throughout to maintain consistency? While the one-way nesting approach and the 3 km resolution evaluation could support the argument that the findings are robust in the second setup, this inconsistency should be better justified. The same applies to the differences in the simulation periods for each run.
- The authors should emphasize more on discussing how the model physics influence the outcomes of the experiments, rather than focusing solely on the statistical evaluation.
- The use of a adaptive timestep can result in unmonitored adjustments, potentially introducing differences in the model simulations and influencing model behavior, particularly under extreme conditions. The authors should provide a justification to ensure that the adaptive timestep does not affect the performance of each member.
- Although there is a clear connection with wind energy resource assessment and applications, the section "Wind Energy Resources" does not align well with the rest of the manuscript and does not provide significant added value. Therefore, it should be removed.
Minor Remarks:
- Line 108: Replace "E2" with "D2-E."
- Table 3: Timestep - Please indicate the exact setup of the adaptive timestep.
- Figures 7, 8, etc.: Include the "light colors/dark colors" distinction in the legend.
Citation: https://doi.org/10.5194/egusphere-2024-3123-RC2 -
CEC1: 'Comment on egusphere-2024-3123 - No compliance with the policy of the journal', Juan Antonio Añel, 07 Dec 2024
reply
Dear authors
Unfortunately, after checking your manuscript, it has come to our attention that it does not comply with our "Code and Data Policy".
https://www.geoscientific-model-development.net/policies/code_and_data_policy.htmlFirst, for model evaluation papers we request that the model name and version number is identified in the title. Therefore, you must modify the title of your manuscript to include this information.
Second, you have archived the WRF code on GitHub. However, GitHub is not a suitable repository for scientific publication. GitHub itself instructs authors to use other long-term archival and publishing alternatives, such as Zenodo.
Third, you state that the code used for data processing and analysis is available upon request. This is clearly forbidden by our policy, which establishes that all the code necessary to replicate the work exposed in a manuscript must be available freely and in one of the repositories that comply with our policy at the submission time. Need to contact authors to get access to code is not acceptable.
Therefore, the current situation with your manuscript is irregular. Please, publish your code in one of the appropriate repositories and reply to this comment with the relevant information (link and a permanent identifier for it (e.g. DOI)) as soon as possible, as we can not accept manuscripts in Discussions that do not comply with our policy. Also, please include the relevant primary input/output data for your simulations.
I have to note that if you do not fix this problem, we will have to reject your manuscript for publication in our journal.
Juan A. Añel
Geosci. Model Dev. Executive EditorCitation: https://doi.org/10.5194/egusphere-2024-3123-CEC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 183 | 58 | 7 | 248 | 3 | 2 |
- HTML: 183
- PDF: 58
- XML: 7
- Total: 248
- BibTeX: 3
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1